At 20:10 22-08-07, you wrote:
>>I built an RX array in 1989 using three sets of crossed dipoles. Each dipole
>>was about 10 feet above the ground and a total of 24 feet long (2 by
>>12 ft pieces
>>of 1" aluminum tubing).
>>One big factor people miss is the bandwidth of the receiver.
Well, not missed by me. BW of the receiver is matched to the BW of
the desired signal, whether it is a DX station or in my case a radar
>Noise power is proportional to receiver bandwidth and a wide
>receiver or detector system limits on propagated noise much more
>easily than a narrow receiver. It doesn't take nearly the receive
>system noise figure to limit on propagated noise when BW is wide.
>For example when I go from 100-250Hz wide CW filters to a 2.1kHz
>wide SSB filter I can drop ten dB out of the system gain and noise
>figure without any problem.
In fact the received wide-band pulse is spread out in the Doppler
domain by the multiplicity
of embedded targets in the D-region. So, in effect, you get a
number of small signals
on different frequencies. Small because the D-region partial
reflections are weak, and a
number of frequencies due to varying Doppler shifts due to the
targets' geometry with respect
to the array.
In effect you have a plurality of narrow receivers, each defined by
the Doppler bin size.
Fortunately for us (and our patents) the phase information for each Doppler bin
is maintained through a complex-FFT, allowing interferometry on each
>Another thing often ignored is the fact signals from the elements in
>an electrically small phased array are not additive in any
>direction. The overall output from the combination of elements is
>less than the output of a single element for both signals and
>external noise. When close spaced elements are combined there not
>only is a reduction in overall output level of signals and noise,
>there is also an even greater reduction of available noise power
>because the array looks at a smaller window of noise.
I have no interest in my application for any signal addition -- each
dipole is handled as an
independent signal and each "target" is located by interferometry
within each corresponding
Doppler bin from all receivers. Neither the antenna nor the receiver
outputs are ever
"added" together. So each Doppler bin is a narrow "receiver" using
a single short-dipole.
Not unlike a narrow receiver, receiving a narrow DX signal, with a
very small antenna and
no preamp. But mathematical beam formation COULD be employed if so
we have the phase of each spectral component. In fact a multiplicity
of beams COULD
be formed simultaneously if so desired.
>Of course local noise floor is also a factor, but most people
>already know the very obvious.
And that is precisely the point. The local noise floor as well as
the atmospheric noise floor
are the defining factors. You have to do your engineering (and you
would agree) in order to
know what you can get away with for system parameters, in any
>This is why, as I clearly stated earlier, there are applications
>where people can get away with a simple transformer. I can here with
>a simple dipole or vertical and with SSB bandwidth, except at the
>most quiet times.
Absolutely. See the above response.
>The point I made, and it is a completely accurate point, is a
>transformer is NOT even remotely close to being the equivalent of a
>voltage follower. The loss of 20-30dB gain following a few dB noise
>figure is significant in many applications, probably more than when
>it isn't significant.
Again, no one is trying to make the claim that a transformer is a
gain stage. It is NOT.
But a transformer is "good enough" in a limited number of
applications -- especially in the
MF band. While your statement is factually true it is not relevant
because we are addressing
an instance (rare as it may or may not be) where low-noise gain is
not required and
may cause additional problems such as IMD.
>There's a big danger in assuming since one particular application
>manages to get away with 20-30 dB less gain and 5-10dB less noise
>figure it generally applies to all other systems. My experience is
>it rarely applies to systems like a compact array used for weak signal DX'ing.
And others have different experiences. That is why the term YMMV was invented.
But if the engineering AND the execution are done properly you can
predict with a high
level of confidence whether the performance will be acceptable.
>I'm sure there are cases where a transformer is fine, but I already
>said that in my very first post.
And we are in total agreement on this point. You DID say that in
some cases a transformer
is just fine and I am just making a point about a specific case (as
is Rick) where you
and Rick and I are all correct. So there is NO DISAGREEMENT.
Fini -- John W0UN
Topband mailing list